这不会节省从数据库加载的时间。但是，你总是可以在数组中取消你不想要的列。我在一个表中有几个列，但不想要一个特别的列。我太懒了，没有把它们都写在SELECT语句中。

$i=0;
$row_array = array();

while($row = mysqli_fetch_assoc($result)){

  $row_array[$i]=$row;
  unset($row_array[$i]['col_name']);
  $i++;
}

DECLARE @SQL VARCHAR(max), @TableName sysname = 'YourTableName'

SELECT @SQL = COALESCE(@SQL + ', ', '') + Name 
FROM sys.columns
WHERE OBJECT_ID = OBJECT_ID(@TableName)
AND name NOT IN ('Not This', 'Or that');

SELECT @SQL = 'SELECT ' + @SQL + ' FROM ' + @TableName

EXEC (@SQL)

更新:

如果您经常使用这个任务，还可以创建一个存储过程来处理它。 在这个例子中，我使用了内置的STRING_SPLIT()，它在SQL Server 2016+上可用， 但如果你需要，有很多例子教你如何在SO上手动创建它。

CREATE PROCEDURE [usp_select_without]
@schema_name sysname = N'dbo',
@table_name sysname,
@list_of_columns_excluded nvarchar(max),
@separator nchar(1) = N','
AS
BEGIN
 DECLARE 
 @SQL nvarchar(max),
 @full_table_name nvarchar(max) = CONCAT(@schema_name, N'.', @table_name);

 SELECT @SQL = COALESCE(@SQL + ', ', '') + QUOTENAME([Name])
 FROM sys.columns sc
 LEFT JOIN STRING_SPLIT(@list_of_columns_excluded, @separator) ss ON sc.[name] = ss.[value]
 WHERE sc.OBJECT_ID = OBJECT_ID(@full_table_name, N'u')
 AND ss.[value] IS NULL;

 SELECT @SQL = N'SELECT ' + @SQL + N' FROM ' + @full_table_name;
 EXEC(@SQL)
END

然后就是:

EXEC [usp_select_without] 
@table_name = N'Test_Table',
@list_of_columns_excluded = N'ID, Date, Name';

你可以创建一个包含你想要选择的列的视图，然后你可以从视图中选择*…

当我使用视图而不是实际表时，BartoszX提出的答案(存储过程)对我不起作用。

下面的想法和代码(除了我的修复)属于BartoszX。

为了使它既适用于表，也适用于视图，使用以下代码:

SET ANSI_NULLS ON
GO

SET QUOTED_IDENTIFIER ON
GO

CREATE PROCEDURE [dbo].[select_without]
@schema_name sysname = N'dbo',
@table_name sysname,
@list_of_columns_excluded nvarchar(max),
@separator nchar(1) = N','
AS
BEGIN
 DECLARE 
 @SQL nvarchar(max),
 @full_table_name nvarchar(max) = CONCAT(@schema_name, N'.', @table_name);

 SELECT @SQL = COALESCE(@SQL + ', ', '') + QUOTENAME([Name])
 FROM sys.columns sc
 LEFT JOIN STRING_SPLIT(@list_of_columns_excluded, @separator) ss ON sc.[name] = ss.[value]
 WHERE sc.OBJECT_ID = OBJECT_ID(@full_table_name)
 AND ss.[value] IS NULL;

 SELECT @SQL = N'SELECT ' + @SQL + N' FROM ' + @full_table_name;
 EXEC(@SQL)
END
GO

是的，这是可能的(但不建议)。

CREATE TABLE contact (contactid int, name varchar(100), dob datetime)
INSERT INTO contact SELECT 1, 'Joe', '1974-01-01'

DECLARE @columns varchar(8000)

SELECT @columns = ISNULL(@columns + ', ','') + QUOTENAME(column_name)
FROM INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_NAME = 'contact' AND COLUMN_NAME <> 'dob'
ORDER BY ORDINAL_POSITION

EXEC ('SELECT ' + @columns + ' FROM contact')

代码说明:

Declare a variable to store a comma separated list of column names. This defaults to NULL. Use a system view to determine the names of the columns in our table. Use SELECT @variable = @variable + ... FROM to concatenate the column names. This type of SELECT does not not return a result set. This is perhaps undocumented behaviour but works in every version of SQL Server. As an alternative you could use SET @variable = (SELECT ... FOR XML PATH('')) to concatenate strings. Use the ISNULL function to prepend a comma only if this is not the first column name. Use the QUOTENAME function to support spaces and punctuation in column names. Use the WHERE clause to hide columns we don't want to see. Use EXEC (@variable), also known as dynamic SQL, to resolve the column names at runtime. This is needed because we don't know the column names at compile time.

这样做不是更简单吗:

sp_help <table_name>

点击“Column_name”列>复制>粘贴(创建一个垂直列表)到新建查询窗口，只需在每个列值前面键入逗号…注释掉你不想要的专栏。比这里提供的任何代码都要少得多，而且仍然是可管理的。